High-precision acoustic detection and recognition based on biomimetic fiber-optic microphones

Acoustic sensing is crucial in nature, supporting functions from biological listening to environmental awareness. In photonics, fiber-optic microphones (FOMs) have become essential tools due to their unique high sensitivity, small size, and resistance to electromagnetic interference, useful in fields such as medical diagnostics, resource exploration, and industrial monitoring. Here, taking advantage of the biomimetic concept, we introduce a fiber-optic acoustic sensing method that offers high-fidelity sound detection and precise voiceprint recognition. Leveraging ingenious design mimicking the sound-sensing membrane of insects, we deeply illustrate how the acoustic response depends on microstructural parameters and develop three types of FOMs with specific micro-electro-mechanical system (MEMS) diaphragms, each with unique spectral responses. Synchronization of the multifrequency response and high sensitivity achieves a detectable pressure from 1.21 to 54.6 μPa / Hz 1/2 across the 20 Hz to 20 kHz band and improves voice recognition accuracy by more than 62.1% with integrated machine learning analysis. Our work integrates micro–nano photonic architecture, fiber-optic sensing microcavities, and deep learning algorithms, offering a scalable solution for intelligent acoustic perception and advancing next-generation optical hearing systems.